INDEX

Explanations

actions or events related to shooting or firing a weapon

oai_token-act-pair · gpt-3.5-turbo

instances of the word "at" used in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alpha

-0.79

Lens

-0.78

ships

-0.77

FTWARE

-0.74

glass

-0.74

hops

-0.74

Cert

-0.73

stars

-0.71

cert

-0.69

NESS

-0.68

POSITIVE LOGITS

 least

1.26

 halftime

0.93

rial

0.92

onement

0.89

abase

0.88

 midday

0.87

 dusk

0.87

roph

0.81

 random

0.79

 midnight

0.79

Activations Density 0.269%

actions or events related to shooting or firing a weapon

instances of the word "at" used in various contexts

No Comments

No Known Activations